Use PiecewiseSequence for List take elements

[danking]

This is just a simplification. List::take already had a specialized implementation. The new PiecewiseSequence specializations in various primitive arrays supersede this change.

EDIT: Heh, I guess the PiecewiseSequence stuff is actually even faster than old List::take. Nice.

---

codex:

Summary

• replace List take's dense per-element index expansion with PiecewiseSequence element runs
• preserve FSL-style null handling: null index/source rows emit zero-length placeholder runs while output validity still comes from Validity::take
• let the elements child choose its own optimized take implementation through the normal ArrayRef::take path

Validation

• cargo +nightly fmt --all
• cargo test -p vortex-array arrays::list::compute::take
• cargo check -p vortex-array
• cargo clippy -p vortex-array --all-targets --all-features -- -D warnings
• git diff --check

[codspeed-hq]

Merging this PR will improve performance by 26.45%

⚡ 13 improved benchmarks
✅ 1785 untouched benchmarks
⏩ 44 skipped benchmarks¹

Performance Changes

	Mode	Benchmark	BASE	HEAD	Efficiency
⚡	Simulation	copy_nullable[65536]	1.4 ms	1 ms	+32.05%
⚡	Simulation	take_filter_list_nullable_random_mask_random_indices[256, 50]	129.3 µs	99.4 µs	+30.05%
⚡	Simulation	take_filter_list_slice_mask_random_indices[768, 50]	119.7 µs	92.4 µs	+29.6%
⚡	Simulation	take_filter_list_nullable_random_mask_random_indices[768, 50]	133.9 µs	103.5 µs	+29.39%
⚡	Simulation	take_filter_list_slice_mask_random_indices[256, 50]	118.8 µs	92 µs	+29.03%
⚡	Simulation	take_filter_list_small_random_mask_random_indices[256, 10]	116.9 µs	90.9 µs	+28.51%
⚡	Simulation	take_filter_list_small_uncached_random_mask_random_indices[256, 10]	117.2 µs	91.6 µs	+28.03%
⚡	Simulation	take_filter_list_small_uncached_random_mask_random_indices[768, 10]	121 µs	94.9 µs	+27.58%
⚡	Simulation	take_filter_list_small_random_mask_random_indices[768, 10]	120.9 µs	94.9 µs	+27.41%
⚡	Simulation	take_filter_list_random_mask_random_indices[256, 50]	121.9 µs	96.1 µs	+26.86%
⚡	Simulation	take_filter_list_random_mask_random_indices[768, 50]	127.2 µs	100.5 µs	+26.55%
⚡	Simulation	copy_non_nullable[65536]	1,089.2 µs	908.6 µs	+19.88%
⚡	Simulation	take_fsl_f16_force_manual_range_copy[1, 10]	14.3 µs	12.9 µs	+10.52%

Tip

Curious why this is faster? Comment @codspeedbot explain why this is faster on this PR, or directly use the CodSpeed MCP with your agent.

---

_{Comparing codex/list-take-piecewise-elements (e0096ca) with develop (add6944)}

Footnotes

1. 44 benchmarks were skipped, so the baseline results were used instead. If they were deleted from the codebase, click here and archive them to remove them from the performance reports. ↩

[robert3005]

Maybe, just maybe this name is too long? OutOffset?

[github-actions]

Polar Signals Profiling Results

Latest Run

Status	Commit	Job	Attempt	Link
🟢 Done	e0096ca		1	Explore Profiling Data

---

Powered by Polar Signals Cloud

[github-actions]

Benchmarks: Vortex queries 📖

Verdict: No clear signal (medium confidence)
Attributed Vortex impact: -4.8%
Engines: DataFusion No clear signal (+9.2%, low confidence) · DuckDB Likely improvement (-17.1%, high confidence)
Vortex (geomean): 0.952x ➖
Parquet (geomean): 0.976x ➖
Shifts: Parquet (control) -2.4% · Median polish -2.4%

How to read Verdict and Engines

• Verdict: Overall PR-level signal after subtracting baseline drift estimated from Parquet control rows. It can be Likely improvement, Likely regression, or No clear signal.
• Engines: Per-engine attribution. DataFusion is compared against DataFusion/Parquet controls; DuckDB is compared against DuckDB/Parquet controls. This answers whether each engine improved or regressed independently.
• Confidence: Based on directional consistency, share of rows above the noise floor, and control-run noise.

---

datafusion / vortex-file-compressed (1.051x ➖, 0↑ 0↓)

name	PR e0096ca (ns)	base add6944 (ns)	ratio (PR/base)
vortex_q00/datafusion:vortex-file-compressed	9786421	9338219	1.05
vortex_q01/datafusion:vortex-file-compressed	6376283	6048928	1.05

datafusion / parquet (0.962x ➖, 0↑ 0↓)

name	PR e0096ca (ns)	base add6944 (ns)	ratio (PR/base)
vortex_q00/datafusion:parquet	20312101	21059503	0.96
vortex_q01/datafusion:parquet	4562846	4750673	0.96

duckdb / vortex-file-compressed (0.821x ✅, 2↑ 0↓)

name	PR e0096ca (ns)	base add6944 (ns)	ratio (PR/base)
vortex_q00/duckdb:vortex-file-compressed 🚀	9936073	12173747	0.82
vortex_q01/duckdb:vortex-file-compressed 🚀	6176471	7472232	0.83

duckdb / parquet (0.991x ➖, 0↑ 0↓)

name	PR e0096ca (ns)	base add6944 (ns)	ratio (PR/base)
vortex_q00/duckdb:parquet	23522457	23276573	1.01
vortex_q01/duckdb:parquet	9398345	9680554	0.97

---

No file size changes detected.